This is an annual workshop to encourage interaction and cooperation among researchers in artificial intelligence with diverse approaches to diagnosis. Previous workshops in this series were held in Val Morin (Canada) in 1996, Goslar (Germany) in 1995, New Paltz (USA) in 1994, Aberystwyth (UK) in 1993, Washington State (USA) in 1992, Milan (Italy) in 1991, Stanford University (USA) in 1990, and in Paris (France) in 1989 (the 0th one).

This year, it will take place in Mont-Saint-Michel, a well-known French site situated at the border between Normandy and Brittany, with four days of presentations and substantial time reserved for discussion. Papers on real-world problem solving (e.g., the field of complex, continuous engineering systems) are especially welcome.

Submissions are welcome on (but not limited to) the following topics:

• theory of diagnosis: abductive, consistency-based, causal, probabilistic, constraint-based, temporal.
• Computational issues: controlling combinatorial explosion, focusing strategies, controlling inference in complex systems, use of structural knowledge, hierarchies.
• Modeling for diagnosis: multiple, approximate, incomplete, probabilistic, functional, and qualitative models, integration of heuristics with model-based diagnosis, principles of modeling, modeling dynamic systems, acquiring models and diagnostic knowledge.
• The diagnosis process: strategies for repair, monitoring, sensor placement, test selection, resource-bounded diagnosis, real-time diagnosis, on-board fault protection.
• Interesting connections between diagnosis and other areas, as traditional FDI techniques, control theory, logic programming, machine learning, and software debugging/synthesis.
• Principled Applications: Of particular interest is the relationship between the techniques applied in practice and computational models of diagnosis. Real-world applications are encouraged from a wide range of fields, such as medicine, chemical, mechanical, electrical, and electronics systems.

Papers should be limited to 5000 words.

In keeping with the desires of the DX-community to establish benchmarks and reference problems for evaluating the current capabilities of diagnosis systems and defining a road map for the next generation of these systems, DX-96 introduced a new track focused on the presentation and discussion of  Reference Problems in Diagnosis which will be present also in DX-97. The objective is to use this as a forum for collecting comprehensive and complete (including test data) problem descriptions that allow other researchers to develop solutions to the problem and perform comparative evaluations. Participants are requested to submit  as a paper (maximum 5000 words) a description of a Reference Problem.

To ensure that the reference problem will make a definitive contribution to the community and the field, we suggest that the paper include the following:

• motivation for the problem
• description of the physical system
• application area and domain(s) covered
• description of primary problem dimensions
• unique characteristics of the problem (e.g., intermittent faults, changing topology, real-time solutions)
• measurements and available data, their source and their quality
• current work on modeling, developing diagnosis algorithms, description and evaluation of solutions
• examples of data and corresponding diagnoses.

Pointers to further details on application area, problem modeling approaches, diagnosis methodologies, and data for evaluation should be provided in terms of literature references and possible Web sites from which this information can be obtained. Note that the papers associated with accepted problems will be published in the Workshop proceedings.

After presentation of the selected reference problems, there will be a special panel session that discusses these problems and attempts to lay a framework for developing future reference problems. The long-term goal of this community is to build a repository of reference problems (cf. the Machine Learning repository of data sets) that would be freely available to diagnosis researchers and help them in achieving the above-mentioned goals. The reference problems may also provide the medium by which real-world practitioners may more closely interact with researchers in the field.

In both categories, please submit four paper copies to Marie-Odile Cordier, the workshop co-chair,  by  May 1st, 1997 at the address indicated below. The additional electronic submission of a postscript file is encouraged.

The papers will be refereed by at least two members of the program committee. Please include postal addresses, electronic mail, fax and telephone numbers on the cover page of all papers.

Authors will be notified of acceptance or rejection by July 8th, 1997. Accepted papers should be revised to accommodate the referee comments before final submission for inclusion in the workshop working notes.

Camera-ready copies of the final paper are due by August 8th, 1997.

For those who wish to attend the Workshop without submitting a paper, please send a short abstract describing your research interests and past research work in the field of diagnosis to Marie-Odile Cordier by May 1st. Invitations will be mailed out by July 8th. To promote active discussion at the Workshop, attendance will be by invitation only.

DX will support student attendance. The grants will cover registration fee, accomodation and meals during the workshop. Should funds allow it, a contribution for travel expenses will be given. Students wishing to get this support must submit to Marie-Odile Cordier by May 1st a one-page statement of interest with a description of their activities in the field and a certificate proving that they are students.

Marie-Odile Cordier
IRISA, Campus de Beaulieu
35042 Rennes Cedex, France
tel: (33) 299847314
fax: (33) 299847171
email: cordier@irisa.fr

Luca Console
Dept of Computer Science, Univ. of Torino
Corso Svizzera 185, 10149 Torino Italy
tel: (39)117429205
fax: (39)11751603
email: lconsole@di.unito.it

Philippe Dague
Univ. Paris-Nord, LIPN
93430 Villetaneuse, France
tel: (33) 149403617
fax: (33) 148260712
email: dague@ura1507.univ-paris13.fr

Sue Abu-Hakima, NRC, Canada
Gautam Biswas, Vanderbilt Univ., USA
Claudia Bottcher, Fraunhofer Inst., Germany
Mike Chantler, Heriot-Watt Univ., Scotland, UK
Luca Chittaro, Univ. of Udine, Italy
Luca Console, Univ. of Torino, Italy
Marie-Odile Cordier, IRISA/Univ. of Rennes, France
Philippe Dague, Univ. Paris-Nord, France
Kourosh Danai, Univ. of Massachusetts, USA
Boi Faltings, EPFL, Switzerland
John Hunt, Univ. of Wales, UK
Sheila Mc Ilraith, Univ. of Toronto, Canada
Jackie Montmain, CEA Marcoule, France
Rob Milne, Intelligent Applic., Scotland, UK
Amit Misra, Vanderbilt Univ., USA
Pandurang Nayak, NASA Ames, USA
Chris Price, Univ. of Wales, UK
Greg Provan, Rockwell, USA
Peter Struss, TU of Munich, Germany
Patrick Taillibert, Dassault Electronique, France
Mugur Tatar, Daimler-Benz, Germany
Daniele Theseider Dupre, Univ. of Torino, Italy
Lyle Ungar, Univ. of Pennsylvania, USA
Frank Van Harmelen,Vrije Univ. Amsterdam, Holland
Takashi Washio, Osaka Univ., Japan
Brian Williams, NASA Ames, USA